Advanced Usage

Extend gs2txt for custom use cases.

Topics

Custom LLM Providers

Create providers for any LLM service:

• Local models (Ollama, vLLM)
• Custom API endpoints
• Retry logic and caching
• Logging and monitoring

Custom Prompts

Customize prompts for specific domains:

• Cancer genomics
• Immunology
• Drug response
• Multi-language support
• Structured output

Custom Enrichment

Implement custom enrichment methods:

• External APIs (DAVID, Enrichr)
• Local databases
• Combined sources
• Pre-computed results

Quick Examples

Custom Provider

from gs2txt.llm.base import BaseLLMProvider

class MyProvider(BaseLLMProvider):
    def generate(self, messages):
        # Your implementation
        return response_text

Custom Prompt

from gs2txt.prompts.builder import PromptBuilder

builder = PromptBuilder(
    system_template="You are a cancer expert...",
    user_template="Analyze: {genes}"
)

Custom Enrichment

from gs2txt.enrichment import BaseEnrichment

class MyEnrichment(BaseEnrichment):
    def enrich(self, genes, **kwargs):
        # Your implementation
        return pd.DataFrame({
            "Term": [...],
            "Adjusted P-value": [...]
        })